An evidence accumulation model of perceptual discrimination with naturalistic stimuli.

Abstract

Evidence accumulation models have been used to describe the cognitive processes underlying performance in tasks involving two-choice decisions about unidimensional stimuli, such as motion or orientation. Given the multidimensionality of natural stimuli, however, we might expect qualitatively different patterns of evidence accumulation in more applied perceptual tasks. One domain that relies heavily on human decisions about complex natural stimuli is fingerprint discrimination. We know little about the ability of evidence accumulation models to account for the dynamic decision process of a fingerprint examiner resolving if two different prints belong to the same finger or not. Here, we apply a dynamic decision-making model-the linear ballistic accumulator (LBA)-to fingerprint discrimination decisions in order to gain insight into the cognitive processes underlying these complex perceptual judgments. Across three experiments, we show that the LBA provides an accurate description of the fingerprint discrimination decision process with manipulations in visual noise, speedaccuracy emphasis, and training. Our results demonstrate that the LBA is a promising model for furthering our understanding of applied decision-making with naturally varying visual stimuli.

What mechanisms mediate prior probability effects on rapid-choice decision-making?

“…However, in recent times, it has been demonstrated that the effect of the speed-accuracy tradeoff is nonselective. Specifically, this non-selectivity of decision-based processes is observed in the quality of sensory evidence (Donkin et al, 2014;Evans, 2021;Palada et al, 2020;Rae et al, 2014;Rafiei & Rahnev, 2021;Sewell & Stallman, 2019;van Maanen et al, 2015;Weigard et al, 2019) and, relevant to the current study, in nondecision-based processes as well, with speed emphasis reducing the time for these processes to complete (Dambacher & Hübner, 2014;Lerche & Voss, 2017;Voss et al, 2004).…”

Evidence accumulation modelling in the wild: understanding safety-critical decisions

“…Finally, parameters controlling between-trial variability in accumulation rate and starting point account for commonly observed differences in the relative speed of correct and incorrect responses [37,38]. Although less commonly interpreted than accumulation rate, threshold, and nondecision time, some studies use variability parameters to identify task factors that lead to increased uncertainty (greater variability) in decision making [26,27,39].…”

“…Recent work has applied EAMs to forensic and medical image discrimination using highly complex naturalistic stimuli encountered in the field (e.g., forensic fingerprint images [39] and histological cell images [58,59]). One study [39] varied the amount of visual noise that was added to naturalistic fingerprint images in an image discrimination task (deciding whether a crime scene print matches a suspect).…”

“…Recent work has applied EAMs to forensic and medical image discrimination using highly complex naturalistic stimuli encountered in the field (e.g., forensic fingerprint images [39] and histological cell images [58,59]). One study [39] varied the amount of visual noise that was added to naturalistic fingerprint images in an image discrimination task (deciding whether a crime scene print matches a suspect). Matching prints were processed less efficiently (with lower accumulation rates) when degraded by visual noise (Figure 2C), which produced more frequent errors, and this deficit was ameliorated by a brief training intervention.…”

“…In terms of managing time pressure, every study in this review that increased time pressure found that individuals proactively set lower thresholds to decrease RT and ensure responses were executed on time [23][24][25][26][27]39,59,61,62]. Individuals adopt this strategy regardless of whether increased time pressure takes the form of tighter deadlines [23][24][25][26]62], being required to process more stimuli per unit time [25][26][27] or being instructed to prioritise speed over accuracy [24,39,59,61], consistent with thresholds as a general mechanism for controlling speedaccuracy trade-offs [7,63]. However, setting thresholds too low can cause unacceptably high error rates and, thus, may be an undesirable strategy in operational settings in which errors are extremely costly (e.g., an air-traffic controller quickly misclassifying aircraft heading for a conflict as safely separated).…”